Mindd Technologies is hiring a junior-level AI/ML Engineer (0–2 years) to own the data and model layer behind our company-intelligence and matchmaking products. You will build the pipelines that feed our models, train and evaluate those models, and ship recommender systems that connect the right companies, investors and users at the right moment.
This is a hands-on role on a small team: you will see your work go from raw data to production ranking in weeks, not quarters.
What you'll do
- Design, build and maintain batch and streaming data pipelines that ingest, clean, enrich and score company and user data from internal and third-party sources.
- Own data quality end to end: schema design, validation, deduplication, entity resolution and monitoring, so downstream models train on data we trust.
- Train, fine-tune and evaluate models for ranking, classification and scoring, including LLM-based components where they beat classical approaches.
- Build and iterate on recommender systems: candidate generation, feature engineering, ranking models, and offline and online evaluation (A/B tests, interleaving, holdouts).
- Ship models to production: feature stores, model serving, versioning, retraining schedules and drift monitoring.
- Define the metrics that tell us a recommendation or score is good, and instrument the product so we can measure them.
- Collaborate with product and engineering to turn vague questions ("which investors should this startup talk to?") into tractable modeling problems.
- Write clear design docs and reviews.
What we're looking for
- 0–2 years of professional experience building ML or data systems that run in production, with at least one shipped recommender, ranking or personalization system you can walk us through.
- Strong Python and SQL. Comfortable in a modern data stack (MongoDB and PostgreSQL).
- Hands-on model training experience with PyTorch, TensorFlow or scikit-learn, plus gradient-boosted trees (XGBoost/LightGBM) for tabular problems.
- Working knowledge of recommender fundamentals: collaborative filtering, two-tower and embedding-based retrieval, learning-to-rank, cold-start strategies, and the evaluation metrics that go with them (NDCG, recall@k, calibration).
- Experience with data quality tooling and practices: schema enforcement, data contracts, validation frameworks (Great Expectations or equivalent), lineage and observability.
- Cloud experience on Azure or GCP, including containerized deployment (Docker; Kubernetes a plus).
- Clear written and verbal communication. You can explain a modeling tradeoff to a business partner and a pipeline failure to a backend engineer.
- Bachelor's or Master's in Computer Science, Statistics, a related field, or equivalent practical experience.
Nice to have
- Experience with LLM-based pipelines: prompt design, structured extraction, retrieval-augmented generation, or using LLMs as judges or labelers.
- Entity resolution or knowledge-graph work on company, person or firmographic data.
- Vector search in production (FAISS, Weaviate or similar).
- Building ML features that serve a mobile product, including latency-sensitive inference.
- Prior work in fintech, venture, B2B data or marketplace products.
- Early-stage startup experience, or comfort owning a system end to end with limited process.